Help needed with tube simulation by LTSpice

LT spice, Curve Captor, PSUDII and whatever other sims you can think of.

Post Reply
Jeffin90620
Posts: 6
Joined: Mon May 20, 2013 5:50 am

Help needed with tube simulation by LTSpice

Post by Jeffin90620 »

While the ultimate goal is to simulate a differential cascode amplifier circuit, I tried starting with a simple Class A tube amp. However, the simulation isn't working (the gain is either zero, or negligibly far from zero).

Would someone please look at the attached file and tell me what I am doing wrong?


Thanks in advance,

Jeff
Attachments
Simple Tube Amp 1.zip
LTSpice circuit for a simple tube amplifier that doesn't work, plus the symbol and definition I downloaded.
(7.18 KiB) Downloaded 364 times
Ray
Posts: 15
Joined: Sat Oct 13, 2012 7:03 pm
Location: Kansas, USA

Post by Ray »

There are a few things that I recommend you change in this circuit. First, the tube has about 0 volts grid-to-cathode bias, owing to a signal source with a 2 volt DC offset. In a Class A circuit you always want the grid to be negative with respect to the cathode; if not, you will get significant output distortion at typical input levels.

Applying DC to the grid in a small signal application like this also isn't advisable if you include a cathode resistor (R1) since the resistor will develop its own voltage drop, which is generally used to bias the tube. This is known as self-bias. While you can bias the tube by applying a DC voltage directly to the grid (a technique called fixed bias) this voltage is normally negative. So the first thing you need to do is to remove this 2-volt DC offset from the signal source. After this change is made, the tube will be biased appropriately, with about -1.4 volts grid-to-cathode with R1 and R2 as indicated.

The 50k grid resistor will work in a circuit with a very low driving impedance, but grid resistor values of 470k to 1 Megohm are typical for this type of circuit.

There will be a significant DC voltage present at the triode's plate, which you generally do not want to apply to the following stage (unless you are intentionally designing a DC coupled amplifier). So to complete this gain stage, you should include a blocking capacitor to remove the DC component from the output and an explicit load resistor for the tube, selecting a value that would be comparable to the load that the circuit is intended to drive.

I've attached a Version 2 schematic that incorporates these changes. I've also added an optional cathode bypass capacitor C2; with this in place, the gain of the stage is about x45; with the capacitor omitted, the gain is about x30.

Although there is nothing that says you need to test under any particular signal conditions, I find it helpful to use a signal frequency of 1 KHz to start with (you use 400 Hz in your original circuit). As a rule, I also like to see four or five cycles in the output waveform which would equate to 5 msec of data collection for 5 complete cycles at 1 KHz. My Version 2 schematic's transient analysis parameters reflect this approach. This just makes the arithmetic a bit easier.

I hope this is helpful.

Ray
Attachments
Simple Tube Amp 2.asc
Modified Simple Tube Amp
(1.64 KiB) Downloaded 357 times
Jeffin90620
Posts: 6
Joined: Mon May 20, 2013 5:50 am

Post by Jeffin90620 »

Thanks for the assistance.

As it turns out, my particular installation of LTSpice will not use the NH5751 definition in dmtriodep.inc. The gain for even this corrected circuit is still around 0.0008.

A member of the Yahoo LTSpice group provided a different model that was accurate enough for me to get some usable numbers in my application.


Jeff
Post Reply